EP0303846B1

EP0303846B1 - A charge coupled device having a parallel-serial converting portion

Info

Publication number: EP0303846B1
Application number: EP88111714A
Authority: EP
Inventors: Takao Kamata
Original assignee: NEC Corp
Current assignee: NEC Corp
Priority date: 1987-07-21
Filing date: 1988-07-20
Publication date: 1997-03-12
Anticipated expiration: 2008-07-20
Also published as: KR890002896A; DE3855818D1; US4990985A; DE3855818T2; JPS6425473A; KR910007407B1; EP0303846A1

Description

Field of the Invention:

This invention relates to a charge coupled device (hereinafter, referred to a CCD) having a parallel-serial converting portion, and more particularly to a charge coupled device such as an area image sensor having a plurality of vertical CCD shift-register and a horizontal CCD shift-register receiving charges from the vertical CCD shift-registers for parallel-serial conversion.

Background of the Invention:

A CCD area image sensor is formed of a plurality of photo-detector elements arranged in rows and columns, a plurality of vertical CCD registers each disposed in parallel with each/column of the photo-detector elements to transfer charges accumulated in the photo-detector elements and a horizontal CCD registers receiving charges from the vertical CCD registers for parallel-serial conversion of charge information. The photo-detectors accumulate charges in response to irradiation of light information. The accumulated charges are simultaneously transferred to the vertical CCD shift-registers to be further transferred the horizontal CCD shift-register. The horizontal CCD shift-register produces the charge information arranged in series which is obtained in a voltage form at a charge-voltage converter provided at an edge of the horizontal CCD shift-register.
The vertical and horizontal CCD shift-registers are driven by two- or four-phase clock pulses. Especially, the horizontal CCD shift-register receives the charges at portions under transfer gate electrodes supplied with the same phase clock from the vertical shift-register. In a non-interlace scanning of the photo-detector array, since the clock pulse applied to the final stage transfer gate electrodes of the vertical CCD shift-registers which are closest to the horizontal CCD shift-register are held at a grounding potential or the lowest potential during the charge-transfer operation of the horizontal CCD shift-register, the charges can be transferred through the horizontal CCD shift-register without going back to the vertical CCD shift-registers. However, in an interlace scanning, the clock pulse applied to the final stage transfer gate electrode changes its phase at every scanning fields. Acccrdingly, there are scanning fields in which the clock pulse applied to the final stage transfer gate electrode is held at a high potential during the charge transfer operation of the horizontal CCD shift-register. In such case, charges in the horizontal CCD shift-register erroneously go back to the vertical CCD shift-registers.
In order to prevent charges in the horizontal CCD shift-register from going back to the vertical CCD shift-register, the CCD area image sensor in the prior art has a particular final stage electrode in the respective vertical CCD shift-registers. The particular final stage electrode is formed at a position closest to the horizontal CCD shift-register and supplied with a particular pulse. The particular pulse has a high level to allow charges to be transferred from the vertical CCD shift-registers to the horizontal CCD shift-register through the particular final stages. This prior art was proposed in a U.S. Patent No. 3,971,003 granted to Walter F. Kosonocky.
The proposed measure has some drawbacks. Firstly, there is a requirement of a generator of the particular pulse. The CCD area image sensor itself requires many pulses for driving the vertical and horizontal CCD shift-registers, for transferring charges from the photo-detector elements to the vertical CCD shift-registers and for driving other circuit portions such as a charge-voltage converter. The increment of necessary pulses makes the pulse generator more complex. Moreover, if the particular pulse has even a small asynchronism with the operation of the horizontal CCD shift-register, the charges in the horizontal CCD shift-register go back to the vertical CCD shift-register. Therefore, the pulse width of the low level signal in the particular pulse should be designed to be wider than an operating period of the horizontal CCD shift-register for allowing the above-mentioned small asynchronism. This allowance of the pulse width in the particular pulse makes the improvement of high speed operation difficult.
Another measure in the prior art is application of a particular D.C. potential to the final stage transfer gate electrode in stead of the particular pulse. The particular D.C. potential is a half potential of the potential difference between high and low levels of the clock pulses applied to the vertical CCD shift-register. This measure is somewhat effective in the image sensor using buried channel type CCD shift-registers. The buried channel type CCD keeps some potential well when the clock pulse has a grounding potential. The charges to be stored in a potential difference between the particular D.C. potential and the potential well at the grounding level of clock pulse are a maximum value of charges transferable in the horizontal CCD shift-register. This transferable charge value is easily affected from changes in the particular D.C. potential and impurity concentration of the buried channel type CCD and is impossible to be made large. Furthermore, if the horizontal CCD shift-register is driven in a sharp clock pulse, charges in the horizontal CCD shift-register go back to the vertical CCD shift-registers prior to charge-transfer along the horizontal CCD shift-register. This sharp clock easily appears in normal clock pulse as an undershoot which is caused by capacitive couplings between transfer gate electrodes, resulting in a limit in high speed operation.
EP-A-0 185 343 discloses a charge coupled device according to the precharacterizing portion of claim 1. In this CCD the potential well in the horizontal channel region is deepened so as to block charges from flowing back to the "narrow-channel" barrier. This deepening is applied to the horizontal channel region. Although this CCD provides an increased transfer speed in the horizontal channels by reducing charge transfer distances, an additional separate final stage transfer gate electrode will be necessary to prevent charges from flowing back over the "narrow channel barrier" into the vertical channels.

Summary of the Invention

It is an object of the present invention to provide a CCD which can effectively prevent charges from flowing back into the vertical channel regions without providing an separate final stage transfer gate electrode.
It is a further object of the present invention to provide a charge coupled device having parallel and serial transfer sections for a parallel-serial conversion in which final stage transfer gate electrodes are removed from the parallel transfer section.
It is another object of the present invention to provide a charge coupled device having a parallel-serial converting portion which is simple in structure and which is operable in a high speed.
These objects are achieved by a charge coupled device according to claim 1. Claim 2 is related to a further advantagous aspect of the present invention.
The selected gate electrodes cover the barrier region together with the channel region in the second CCD shift-register. Only by applying a clock pulse to the selected gate electrodes, potential barriers against a charge flow for going back to the first CCD shift-registers from the second CCD-register are always formed. Therefore, the particular final stage transfer gate electrodes supplied with a paticular pulse of D.C. potential are not required to simplify a whole structure and operating pulse generator. Furthermore, since the potential barriers are automatically formed by application of clock pulse to the second CCD shift-register, the operation speed of the second CCD shift-register can be increased without charges going back to the first CCD shift registers. Especially, the barrier is automatically formed, if the clock pulse becomes sharp by undershoot caused by capacitive coupling between transfer gates. Thus, the high speed operation is not damaged by the undershoot in clock pulse. Moreover, the transferable charge value through the second CCD shift-register is arbitrarily set by controlling the impurity concentration in the high impurity concentration regions. This value may be made larger, compared to the prior art using the particular D.C. potential.

Brief Description of the Drawings

The above and further objects, features and advantages of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, wherein:

Fig. 1 is a plan view of a principal part of a first preferred embodiment of the present invention;
Fig. 2 is a sectional view taken along the line A-A' in Fig. 1;
Fig. 3 is a diagram showing a change in potential under a transfer gate electrode supplied with a clock pulse φ_H2 in Figs. 1 and 2; and
Fig. 4 is a plan view of a principal part of a second preferred embodiment of the present invention.

Description of the Preferred Embodiment

A parallel-serial converter using CCD shift-registers are shown in Figs. 1, 2 and 3 as the first preferred embodiment of the present invention. A P-type silicon substrate having a resistivity of 10 to 15 ohm·cm is used as a substrate 1. At a surface of the substrate 1, three lines of N-type well regions 16 and a stripe of another N-type well region 15 are formed as channel regions for vertical and horizontal CCD shift-registers of buried channel type. These N-type well regions have an impurity concentration of 1 to 3 x 10¹⁶ cm^-3 and a depth of 0.6 µm, respectively. At the surface of the substrate 1 except for the N- type well regions 15 and 16, a channel stopper region 2 is formed with a P⁺-type impurity concentration of 5 to 10 x 10¹⁷ cm^-3 and a depth of about 0.6 µm. A thin silicon oxide film 17 having a thickness of 60 to 90 nm (600 to 900 Å) is formed to cover the N- type well regions 15 and 16 as a gate insulator film. A thick silicon oxide film 18 having a thickness of about 1 µm is formed over the channel stopper region 2.
The three vertical CCD shift-registers are formed by forming a plurality of transfer gate electrodes 4, 5, 6 and 7 so as to cross the N-type well regions 16 via the thin and thick silicon oxide films 17 and 18. Those transfer gate electrodes 4, 5, 6 and 7 are formed with a two layer structure. The transfer gate electrodes 5 and 7 are lower layer electrodes of polycrystalline silicon covered with silicon oxide film. The other transfer gate electrodes 4 and 6 are upper layer electrodes having parts overlying their adjacent lower layer electrodes 5 and 7 and being made of polycrystalline silicon or aluminium. Those transfer gate electrodes 4, 5, 6 and 7 are supplied with four-phase clock pulses φ_V1' φ_V2' φ_V3 and φ_V4 to simultaneously drive the vertical shift registers.
The horizontal CCD shift-register is formed by disposing transfer gate electrodes 8, 9, 10 and 11 on the N-type well region 15. The transfer gate electrodes 8, 9, 10 and 11 are also formed with two layer structure. Lower layer electrodes 9 and 11 are formed of polycrystalline silicon layer covered with silicon oxide film. By using those lower layer electrodes 9 and 10 as a mask, barrier regions 3 having an N-type impurity concentration of 2 to 5 x 10¹⁵ cm^-3 are formed by ion-implantation of P-type impurities. Similar impurity ion-implantation is carrier out with a mask of the lower layer electrode 7 and 9 to form barrier regions 3a having an N-type impurity concentration of 2 to 5 x 10¹⁵ cm^-3 at connecting portion of the N-type well regions 16 with the N-type well region 15. Upper layer electrodes 8 and 10 are formed on those barrier regions 3 with polycrystalline silicon or aluminium so as to overlie their adjacent lower layer electrodes 9 and 11. The upper layer electrodes 8 are connected with each other by a connecting electrode 18 disposed in parallel with the transfer gate electrode 4, 5, 6 and 7 for the vertical CCD shift-registers. The connecting electrode 18 overlies the barrier regions 3a. The impurity concentration of the barrier region 3a should be selected in accordance with the required charge value transferable through the horizontal shift-register formed along the N-type well region 15.
The horizontal shift register is driven with two-phase clock pulses φ_H1 and φ_H2. The clock pulse φ_H1 is applied to the lower layer electrode 11 and the upper layer electrode 10. The other clock pulse φ_H2 is applied to the lower layer electrode 9 and the upper layer electrode 8 together with the connecting electrode 18.
The change in potential wells at a portion under the lower layer electrode 9 and the barrier region 3a under the connecting electrode 18 will next be explained with reference to Fig. 3. When the clock pulse φ_H2 has a high level potential, the potential under the lower layer electrode 9 becomes to have levels 20 and 21 of the solid line. The stepped level 21 is generated by a narrow channel effect occurring at the narrow N-type regions 16. In the same time, the potential at the barrier regions 3a is as shown by the solid line 22. The charge value in the potential well enclosed by the solid lines 20, 21 and 22 is a transferable charge value. The potential well enclosed by the solid lines 20, 21 and 22 is raised as shown by the dotted lines 20', 21' and 22' with keeping their mutual relation, when the clock pulse φ_H2 is lowered to a lower potential or a grounding potential. The same transferable charge value is kept without going back to the vertical CCD shift-registers, irrespective of any potential level of the clock pulse φ_V4.
The potential barrier shown by the solid line 22 and the dotted line 22' is always generated in accordance with the clock pulse φ_H2 and the impurity concentration in the barrier region 3a. This potential barrier defines the maximum transferable charge value through the horizontal CCD shift-register and functions as a barrier for preventing charges in the horizontal CCD shift-register from going back to the vertical CCD shift-registers. Thus, the final stage transfer gate electrode is not required in the vertical CCD shift-registers. This removal of the final stage electrode allows to disuse the particular pulse or the particular D.C. potential required in the prior art device, resulting in a possibility of high speed operation. This is because the particular pulse requiring a timing allowance for the clock pulses of the horizontal shift-register is not used. Another reason is because a sufficient barrier is formed even if the horizontal shift-register is driven by sharp clock pulses.
The potential barriers of the solid line 22 and the dotted line 22' always keep their potential difference from the potential at the channel region of the solid lines 20 and 21 and the dotted line 20' and 21'. The potential difference can be controlled to have a considerable value by the impurity concentration of the barrier region 3a. In this way, the maximum transferable charge value can be precisely controlled and can be made large, compared to the prior art device using the final stage electrode supplied with the particular D.C. potential.
It is needless to say that the removal of the final stage electrode makes the device structure and the driving pulse generator simple. The barrier regions 3a are formed with an ion-implantation step using a mask of lower layer electrodes 7 and 9 in the same ion-implantation process for forming the barrier regions 3 in the N-type well region 15. Any additional manufacturing steps are not required by the formation of the barrier regions 3a.
The second preferred embodiment of the present invention shown in Fig. 4 is a CCD area image sensor. Major part of the CCD area image sensor is similar to the parallel-serial converter shown in Figs. 1 and 2 and uses the same reference numerals for the same portions to omit their explanations. A plurality of photo-detector elements 30 are inserted between N-type well regions 16. Each photo-detector element 30 is formed of a P-type region formed in an N-type well region (not shown in Fig. 4). The N-type well regions 16 have connecting portions to adjacent photo-detector elements 30. The transfer gate electrodes 4, 5, 6 and 7 have, respectively, many openings to expose the photo-detector elements 30. The transfer gate electrodes 5 and 7 which cover the connecting portions of the N-type well regions 16 to the photo-detector elements, are supplied with a clock pulse superimposed with a charge transfer pulse having a potential higher than a high level potential in a clock pulse for charge transfer through a CCD shift-register. With the charge transfer pulse, charges accumulated in the photo-detector elements in accordance with the irradiated light are transferred to the vertical CCD shift-registers. The transferred charges are parallely shifted along the vertical CCD shift-registers to the horizontal CCD shift-register for parallel-serial conversion.
According to the area image sensor of the present invention, an interlace scanning is easily achieved with a simple device structure and a decreased number of control pulses. Since the interlace scanning can be performed with high speed clock pulses, the number of photo-detector elements integrated in a single image sensor chip can be increased to improve resolution.
Although impurity concentration regions are used as barrier regions for preventing charges in the horizontal shift-register from going back to the vertical shift-registers, such barrier regions may be formed with a thick insulator film. The inventive feature of the present invention may be applicable to a parallel-serial converter using surface channel type CCD, in stead of buried channel type CCD used in the above-mentioned embodiment of the present invention. In the case using the surface channel type CCD's, the barrier regions should be the same conductivity type regions as the substrate but should have an impurity concentration higher than the substrate.

Claims

A charge coupled device comprising a plurality of first CCD shift-registers (4-7, 16) transferring charges in parallel, each of said first CCD shift-registers including a first channel region (16) insulatingly covered with first transfer gate electrodes (4-7), and
a second CCD shift-register (8-11, 15) receiving said charges from said first CCD shift-registers for a parallel-serial conversion and including a second channel region (15) insulatingly covered with second transfer gate electrodes (8-11), said second CCD shift-register being connected to said first CCD shift-registers through narrow channel barrier regions (21) occurring at said narrow first channel regions and covered with transfer gate electrodes (9) of said second CCD shift-register,
characterized by
further comprising a plurality of additional barrier regions (3a) each formed between each of said first channel regions and an associated one of said narrow channel barrier regions and insulatingly covered with a common transfer gate electrode (8, 18) of said second CCD shift-register so as to block charges from flowing back from the second channel region over said narrow channel barrier regions into said first channel regions.
The charge coupled device as claimed in claim 1, characterized in that it is of buried channel type and said additional barrier regions (3a) have the same conductivity type as said first channel regions (16) and an impurity concentration lower than said first channel regions (16).